Real time emotion recognition

• Main Author: Cheang, Khai Mun
• Other Authors: Smitha Kavallur Pisharath Gopi
• Format: Final Year Project
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/70443
• Institution: Nanyang Technological University

Recognizing emotion from facial expression is a fundamental aspect of interpersonal communication. Children with diseases like autism spectrum disorder, Asperger’s Syndrome, often face difficulties trying to understand other people’s emotions. Consequently, such children need to be taught explicitly how to read other people’s mood from nonverbal communication channels such as the facial expressions. For this project, I will try to estimate real time emotion recognition of 6 basic emotion proposed by Ekman (1972) [1]: Anger, Disgust, Happy, Sad and surprise plus one additional neutral emotion from input videos. This paper will focus on comparing different method of feature extractor and machine learning algorithms and implement the most suitable method in real time emotion recognition. The facial expression recognition software is written in python, with machine learning library like OpenCV, Scikit-learn. The proposed method has achieved 87.78% of accuracy with 7 emotions.